نبذة مختصرة : The efficiency of Monte-Carlo procedures to test some hypotheses about the spatial patterns of larvae and damages of Lobesia botrana was studied. Two hypotheses were tested to detect spatial heterogeneity and spatial dependence. The most practical implication is to provide an efficient sampling scheme. The study of the relationship between spatial patterns and grape availability was required to explain scales of spatial heterogeneity and population dynamics studies were needed to relate it to oviposition behavior. It was tested through a third hypothesis. We adapted Monte-Carlo simulation procedures for the analysis of exhaustive count data obtained from regular grids delimited within each of two vineyards. Statistical analyses were based on count permutations and on count redistributions according to the hypotheses which were tested. Indices of aggregation and autocorrelation statistics were used. The hypotheses that we tested at different scales were random distribution of the infestations (HR), independence of vine stock (or groups of k vine stocks) infestation (HI) and independence between vine stock infestation and grape availability (HG). Monte-Carlo tests revealed the same spatial patterns for larvae and damages. We detected different spatial patterns. The implications for sampling were that sample unit could be an individual stock and that sampling along a row could not be used to estimate population density in the vineyard. Results showed that infestation of a given stock depended on grape availability on this stock and on neighboring vine stocks. ; Isabelle Badenhausser, Patrice Lecharpentier, Lionel Delbac, Pascale Pracros ; Lit
No Comments.